Abstract P111

KEYFORM-008: Randomized Phase 2 Study of Coformulated Favezelimab/Pembrolizumab Versus Chemotherapy in Participants With Relapsed or Refractory Classic Hodgkin Lymphoma

Background: PD-1 blockade (eg, pembrolizumab) is standard treatment for relapsed or refractory (R/R) classic Hodgkin lymphoma (cHL), but most patients experience progressive disease (PD), and optimal post–PD-1 therapy is undefined. Favezelimab (anti-LAG3) plus pembrolizumab showed promising activity and manageable safety in participants (pts) with anti–PD-1–refractory R/R cHL in a phase 1/2 study. We report results from KEYFORM-008 (NCT05508867), a randomized, open-label, phase 2 study of coformulated favezelimab/pembrolizumab vs chemotherapy in anti–PD-1–refractory R/R cHL.

Methods: Adults with R/R cHL, ECOG PS 0-2, and PD per investigator ≤12 weeks from last dose of anti–PD-(L)1–based therapy (received for ≥3 months with ≥2 doses) were randomly assigned 1:1 to favezelimab 800 mg/pembrolizumab 200 mg IV Q3W or chemotherapy (gemcitabine 800-1200 mg/m2 IV on days 1 and 8 of a 21-day cycle or bendamustine 90-120 mg/m2 IV on days 1 and 2 of a 21- or 28-day cycle). Prior brentuximab vedotin and autologous stem cell transplant were required. Crossover to favezelimab/pembrolizumab was allowed for pts who received chemotherapy with PD per Lugano criteria by BICR. Primary end point was PFS per Lugano criteria by investigator. Secondary end points included OS, ORR, and DOR per Lugano criteria by investigator, and safety. Primary hypothesis for PFS was evaluated using a stratified log-rank test and at 5% (1-sided) alpha level.

Results: Overall, 203 pts were randomly assigned to favezelimab/pembrolizumab (n=104) and chemotherapy (n=99); 52 pts in the chemotherapy arm crossed over, with 17 ongoing treatment. At data cutoff (Sept 3, 2025), median follow-up was 17.6 months (range, 7.9-34.3). Efficacy data are reported in the table. Treatment-related AEs (TRAEs) occurred in 79 pts (76.0%) with favezelimab/pembrolizumab and 81 pts (83.5%) with chemotherapy. Grade 3-5 TRAEs occurred in 19 (18.3%) and 39 pts (40.2%). One treatment-related death occurred in the chemotherapy arm only (sepsis).

Summary/Conclusion: Primary end point was met, with improved PFS and late curve separation at 6 months, although few pts at risk thereafter warrant caution. DOR was longer; OS trended favorably but remains immature. ORR was lower for favezelimab/pembrolizumab, but disease control rates (CR+PR+SD) were similar between arms. KEYFORM-008 was discontinued due to sponsor’s business decision to end the entire favezelimab clinical development program and was unrelated to safety or efficacy.